FS-UAE for OS X PPC (Not officially supported)

[SnakeCoils]

PPC Macs are still strong but not enough to run smoothly an accurate emulator like FS-UAE with the layer of an emulated PPC Amiga board. It could be enabled thanks to a workaround (by tlosm) that allow the QEMU core to run so OS4 but as long you have an hi-end quad-core G5 the results are quite disappointing on all other PowerMacs.
However, if you want to give it a try I can prepare a modified binary you can play with, I hope the tlosm patch will apply to latest dev releases.

[Marmotta]

I have G5 2.3 dual core... I have tried some tlosm packages but I not able to start emulator. My last error is relative to qemu.

[SnakeCoils]

The last 2.7.13dev binary for OSX PPC 10.5.8 is available in The Zone!

Same issue as latest dev release: no text string visible in the Cmd+F12 menu.
Frode, could you please look if the osd code is still big-endian friendly? Many thanks :-)

[FrodeSolheim]

Originally Posted by SnakeCoils

Frode, could you please look if the osd code is still big-endian friendly? Many thanks :-)

But I wouldn't introduce Big-endian-only bugs on purpose. I can't really know if there is a bug or not without debugging on a PPC Mac, which I am not going to do.

Perhaps I increased the OpenGL texture size (or something like that) a while ago, and that broke rendering on your computer if max texture size is exceeded, or something else along those lines.

You should try to find the exact commit (or at least minimum version) where the problem first started occuring, then we might be able to reason about what the problem can be.

[SnakeCoils]

Well, I have searched in my offline and online archive along some re-build of older FS-UAE version, I can tell for sure that menu text strings were visible until 2.5.25dev and no more visible on 2.7.5dev version.
I have tried to search in your online archive the latest 2.5.41dev source for further narrowing the investigation but seems the whole 2.5.x dev cycle is missing (or, more likely, I am looking in the wrong place).
The 2.7.0-2.7.4 range report build errors and the 2.7.5 is the first to be friendly with OSX PPC again.
If you redirect me when I can still download the 2.5.x sources again I will experiment further reporting the issue as soon as possible.

[FrodeSolheim]

2.5.x version were removed due to storage capacity issues. You can find them here:
https://www.dropbox.com/sh/1vuodhdfm...tRga5fMWa?dl=0

[SnakeCoils]

Originally Posted by FrodeSolheim

2.5.x version were removed due to storage capacity issues. You can find them here:
https://www.dropbox.com/sh/1vuodhdfm...tRga5fMWa?dl=0

Just looked in but I found there only a "static" folder containing 32 and 64 bit executables for Windows.

[FrodeSolheim]

Oops, wrong like, this is the correct:
https://www.dropbox.com/sh/8sc0g1of8...ZRhooKEka?dl=0

[SnakeCoils]

Thanks Frode :-) Now I am only waiting my G4 to finish the MacPorts update and then I can start to experiment. The libgcc update started over 20 hours ago and still going on (and no, the PowerMac is not freezed...) :-(

[SnakeCoils]

I'm trying in re-compile some old sources but I am stopped by errors I have not seen at the time, like this from 2.6.2 stable release:

libtool: compile:  gcc-mp-5 -DHAVE_CONFIG_H -I. -I../include -I../include -I../include -Wall -Wpointer-arith -Wcast-align -Wcast-qual -Wstrict-prototypes -Wshadow -Waggregate-return -Wmissing-prototypes -Wnested-externs -Wsign-compare -g -maltivec -mabi=altivec -O3 -fomit-frame-pointer -fno-common -std=gnu89 -faltivec -g -O2 -maltivec -mabi=altivec -MT libmpeg2arch_la-motion_comp_altivec.lo -MD -MP -MF .deps/libmpeg2arch_la-motion_comp_altivec.Tpo -c motion_comp_altivec.c  -fno-common -DPIC -o .libs/libmpeg2arch_la-motion_comp_altivec.o
motion_comp_altivec.c: In function 'my_vec_ld':
motion_comp_altivec.c:48:23: warning: cast discards 'const' qualifier from pointer target type [-Wcast-qual]
     return vec_ld (A, (uint8_t *)B);
                       ^
libtool: compile:  gcc-mp-5 -DHAVE_CONFIG_H -I. -I../include -I../include -I../include -Wall -Wpointer-arith -Wcast-align -Wcast-qual -Wstrict-prototypes -Wshadow -Waggregate-return -Wmissing-prototypes -Wnested-externs -Wsign-compare -g -maltivec -mabi=altivec -O3 -fomit-frame-pointer -fno-common -std=gnu89 -faltivec -g -O2 -maltivec -mabi=altivec -MT libmpeg2arch_la-motion_comp_altivec.lo -MD -MP -MF .deps/libmpeg2arch_la-motion_comp_altivec.Tpo -c motion_comp_altivec.c -o libmpeg2arch_la-motion_comp_altivec.o >/dev/null 2>&1
mv -f .deps/libmpeg2arch_la-motion_comp_altivec.Tpo .deps/libmpeg2arch_la-motion_comp_altivec.Plo
/bin/sh ../libtool  --tag=CC   --mode=compile gcc-mp-5 -DHAVE_CONFIG_H -I. -I../include  -I../include -I../include  -Wall -Wpointer-arith -Wcast-align -Wcast-qual -Wstrict-prototypes -Wshadow -Waggregate-return -Wmissing-prototypes -Wnested-externs -Wsign-compare -g  -maltivec -mabi=altivec -O3 -fomit-frame-pointer -fno-common -std=gnu89 -faltivec  -g -O2 -maltivec -mabi=altivec -MT libmpeg2arch_la-idct_altivec.lo -MD -MP -MF .deps/libmpeg2arch_la-idct_altivec.Tpo -c -o libmpeg2arch_la-idct_altivec.lo `test -f 'idct_altivec.c' || echo './'`idct_altivec.c
libtool: compile:  gcc-mp-5 -DHAVE_CONFIG_H -I. -I../include -I../include -I../include -Wall -Wpointer-arith -Wcast-align -Wcast-qual -Wstrict-prototypes -Wshadow -Waggregate-return -Wmissing-prototypes -Wnested-externs -Wsign-compare -g -maltivec -mabi=altivec -O3 -fomit-frame-pointer -fno-common -std=gnu89 -faltivec -g -O2 -maltivec -mabi=altivec -MT libmpeg2arch_la-idct_altivec.lo -MD -MP -MF .deps/libmpeg2arch_la-idct_altivec.Tpo -c idct_altivec.c  -fno-common -DPIC -o .libs/libmpeg2arch_la-idct_altivec.o
idct_altivec.c:62:51: warning: left-hand operand of comma expression has no effect [-Wunused-value]
 #define VEC_S16(a,b,c,d,e,f,g,h) (vector_s16_t) (a, b, c, d, e, f, g, h)
                                                   ^
idct_altivec.c:66:5: note: in expansion of macro 'VEC_S16'
     VEC_S16 (23170, 13573, 6518, 21895, -23170, -21895, 32, 31);
     ^
idct_altivec.c:62:54: warning: left-hand operand of comma expression has no effect [-Wunused-value]
 #define VEC_S16(a,b,c,d,e,f,g,h) (vector_s16_t) (a, b, c, d, e, f, g, h)
                                                      ^
idct_altivec.c:66:5: note: in expansion of macro 'VEC_S16'
     VEC_S16 (23170, 13573, 6518, 21895, -23170, -21895, 32, 31);
     ^
idct_altivec.c:62:57: warning: left-hand operand of comma expression has no effect [-Wunused-value]
 #define VEC_S16(a,b,c,d,e,f,g,h) (vector_s16_t) (a, b, c, d, e, f, g, h)
                                                         ^
idct_altivec.c:66:5: note: in expansion of macro 'VEC_S16'
     VEC_S16 (23170, 13573, 6518, 21895, -23170, -21895, 32, 31);
     ^
idct_altivec.c:62:60: warning: left-hand operand of comma expression has no effect [-Wunused-value]
 #define VEC_S16(a,b,c,d,e,f,g,h) (vector_s16_t) (a, b, c, d, e, f, g, h)
                                                            ^
idct_altivec.c:66:5: note: in expansion of macro 'VEC_S16'
     VEC_S16 (23170, 13573, 6518, 21895, -23170, -21895, 32, 31);
     ^
idct_altivec.c:62:63: warning: left-hand operand of comma expression has no effect [-Wunused-value]
 #define VEC_S16(a,b,c,d,e,f,g,h) (vector_s16_t) (a, b, c, d, e, f, g, h)
                                                               ^
idct_altivec.c:66:5: note: in expansion of macro 'VEC_S16'
     VEC_S16 (23170, 13573, 6518, 21895, -23170, -21895, 32, 31);
     ^
idct_altivec.c:62:66: warning: left-hand operand of comma expression has no effect [-Wunused-value]
 #define VEC_S16(a,b,c,d,e,f,g,h) (vector_s16_t) (a, b, c, d, e, f, g, h)
                                                                  ^
idct_altivec.c:66:5: note: in expansion of macro 'VEC_S16'
     VEC_S16 (23170, 13573, 6518, 21895, -23170, -21895, 32, 31);
     ^
idct_altivec.c:62:69: warning: left-hand operand of comma expression has no effect [-Wunused-value]
 #define VEC_S16(a,b,c,d,e,f,g,h) (vector_s16_t) (a, b, c, d, e, f, g, h)
                                                                     ^
idct_altivec.c:66:5: note: in expansion of macro 'VEC_S16'
     VEC_S16 (23170, 13573, 6518, 21895, -23170, -21895, 32, 31);
     ^
idct_altivec.c:66:5: error: can't convert between vector values of different size
idct_altivec.c:62:51: warning: left-hand operand of comma expression has no effect [-Wunused-value]
 #define VEC_S16(a,b,c,d,e,f,g,h) (vector_s16_t) (a, b, c, d, e, f, g, h)
                                                   ^
idct_altivec.c:68:5: note: in expansion of macro 'VEC_S16'
     VEC_S16 (16384, 22725, 21407, 19266, 16384, 19266, 21407, 22725);
     ^
idct_altivec.c:62:54: warning: left-hand operand of comma expression has no effect [-Wunused-value]
 #define VEC_S16(a,b,c,d,e,f,g,h) (vector_s16_t) (a, b, c, d, e, f, g, h)
                                                      ^
idct_altivec.c:68:5: note: in expansion of macro 'VEC_S16'
     VEC_S16 (16384, 22725, 21407, 19266, 16384, 19266, 21407, 22725);
     ^
idct_altivec.c:62:57: warning: left-hand operand of comma expression has no effect [-Wunused-value]
 #define VEC_S16(a,b,c,d,e,f,g,h) (vector_s16_t) (a, b, c, d, e, f, g, h)
                                                         ^
idct_altivec.c:68:5: note: in expansion of macro 'VEC_S16'
     VEC_S16 (16384, 22725, 21407, 19266, 16384, 19266, 21407, 22725);
     ^
idct_altivec.c:62:60: warning: left-hand operand of comma expression has no effect [-Wunused-value]
 #define VEC_S16(a,b,c,d,e,f,g,h) (vector_s16_t) (a, b, c, d, e, f, g, h)
                                                            ^
idct_altivec.c:68:5: note: in expansion of macro 'VEC_S16'
     VEC_S16 (16384, 22725, 21407, 19266, 16384, 19266, 21407, 22725);
     ^
idct_altivec.c:62:63: warning: left-hand operand of comma expression has no effect [-Wunused-value]
 #define VEC_S16(a,b,c,d,e,f,g,h) (vector_s16_t) (a, b, c, d, e, f, g, h)
                                                               ^
idct_altivec.c:68:5: note: in expansion of macro 'VEC_S16'
     VEC_S16 (16384, 22725, 21407, 19266, 16384, 19266, 21407, 22725);
     ^
idct_altivec.c:62:66: warning: left-hand operand of comma expression has no effect [-Wunused-value]
 #define VEC_S16(a,b,c,d,e,f,g,h) (vector_s16_t) (a, b, c, d, e, f, g, h)
                                                                  ^
idct_altivec.c:68:5: note: in expansion of macro 'VEC_S16'
     VEC_S16 (16384, 22725, 21407, 19266, 16384, 19266, 21407, 22725);
     ^
idct_altivec.c:62:69: warning: left-hand operand of comma expression has no effect [-Wunused-value]
 #define VEC_S16(a,b,c,d,e,f,g,h) (vector_s16_t) (a, b, c, d, e, f, g, h)
                                                                     ^
idct_altivec.c:68:5: note: in expansion of macro 'VEC_S16'
     VEC_S16 (16384, 22725, 21407, 19266, 16384, 19266, 21407, 22725);
     ^
idct_altivec.c:68:5: error: can't convert between vector values of different size
idct_altivec.c:62:51: warning: left-hand operand of comma expression has no effect [-Wunused-value]
 #define VEC_S16(a,b,c,d,e,f,g,h) (vector_s16_t) (a, b, c, d, e, f, g, h)
                                                   ^
idct_altivec.c:70:5: note: in expansion of macro 'VEC_S16'
     VEC_S16 (16069, 22289, 20995, 18895, 16069, 18895, 20995, 22289);
     ^
idct_altivec.c:62:54: warning: left-hand operand of comma expression has no effect [-Wunused-value]
 #define VEC_S16(a,b,c,d,e,f,g,h) (vector_s16_t) (a, b, c, d, e, f, g, h)
                                                      ^
idct_altivec.c:70:5: note: in expansion of macro 'VEC_S16'
     VEC_S16 (16069, 22289, 20995, 18895, 16069, 18895, 20995, 22289);
     ^
idct_altivec.c:62:57: warning: left-hand operand of comma expression has no effect [-Wunused-value]
 #define VEC_S16(a,b,c,d,e,f,g,h) (vector_s16_t) (a, b, c, d, e, f, g, h)
                                                         ^
idct_altivec.c:70:5: note: in expansion of macro 'VEC_S16'
     VEC_S16 (16069, 22289, 20995, 18895, 16069, 18895, 20995, 22289);
     ^
idct_altivec.c:62:60: warning: left-hand operand of comma expression has no effect [-Wunused-value]
 #define VEC_S16(a,b,c,d,e,f,g,h) (vector_s16_t) (a, b, c, d, e, f, g, h)
                                                            ^
idct_altivec.c:70:5: note: in expansion of macro 'VEC_S16'
     VEC_S16 (16069, 22289, 20995, 18895, 16069, 18895, 20995, 22289);
     ^
idct_altivec.c:62:63: warning: left-hand operand of comma expression has no effect [-Wunused-value]
 #define VEC_S16(a,b,c,d,e,f,g,h) (vector_s16_t) (a, b, c, d, e, f, g, h)
                                                               ^
idct_altivec.c:70:5: note: in expansion of macro 'VEC_S16'
     VEC_S16 (16069, 22289, 20995, 18895, 16069, 18895, 20995, 22289);
     ^
idct_altivec.c:62:66: warning: left-hand operand of comma expression has no effect [-Wunused-value]
 #define VEC_S16(a,b,c,d,e,f,g,h) (vector_s16_t) (a, b, c, d, e, f, g, h)
                                                                  ^
idct_altivec.c:70:5: note: in expansion of macro 'VEC_S16'
     VEC_S16 (16069, 22289, 20995, 18895, 16069, 18895, 20995, 22289);
     ^
idct_altivec.c:62:69: warning: left-hand operand of comma expression has no effect [-Wunused-value]
 #define VEC_S16(a,b,c,d,e,f,g,h) (vector_s16_t) (a, b, c, d, e, f, g, h)
                                                                     ^
idct_altivec.c:70:5: note: in expansion of macro 'VEC_S16'
     VEC_S16 (16069, 22289, 20995, 18895, 16069, 18895, 20995, 22289);
     ^
idct_altivec.c:70:5: error: can't convert between vector values of different size
idct_altivec.c:62:51: warning: left-hand operand of comma expression has no effect [-Wunused-value]
 #define VEC_S16(a,b,c,d,e,f,g,h) (vector_s16_t) (a, b, c, d, e, f, g, h)
                                                   ^
idct_altivec.c:72:5: note: in expansion of macro 'VEC_S16'
     VEC_S16 (21407, 29692, 27969, 25172, 21407, 25172, 27969, 29692);
     ^
idct_altivec.c:62:54: warning: left-hand operand of comma expression has no effect [-Wunused-value]
 #define VEC_S16(a,b,c,d,e,f,g,h) (vector_s16_t) (a, b, c, d, e, f, g, h)
                                                      ^
idct_altivec.c:72:5: note: in expansion of macro 'VEC_S16'
     VEC_S16 (21407, 29692, 27969, 25172, 21407, 25172, 27969, 29692);
     ^
idct_altivec.c:62:57: warning: left-hand operand of comma expression has no effect [-Wunused-value]
 #define VEC_S16(a,b,c,d,e,f,g,h) (vector_s16_t) (a, b, c, d, e, f, g, h)
                                                         ^
idct_altivec.c:72:5: note: in expansion of macro 'VEC_S16'
     VEC_S16 (21407, 29692, 27969, 25172, 21407, 25172, 27969, 29692);
     ^
idct_altivec.c:62:60: warning: left-hand operand of comma expression has no effect [-Wunused-value]
 #define VEC_S16(a,b,c,d,e,f,g,h) (vector_s16_t) (a, b, c, d, e, f, g, h)
                                                            ^
idct_altivec.c:72:5: note: in expansion of macro 'VEC_S16'
     VEC_S16 (21407, 29692, 27969, 25172, 21407, 25172, 27969, 29692);
     ^
idct_altivec.c:62:63: warning: left-hand operand of comma expression has no effect [-Wunused-value]
 #define VEC_S16(a,b,c,d,e,f,g,h) (vector_s16_t) (a, b, c, d, e, f, g, h)
                                                               ^
idct_altivec.c:72:5: note: in expansion of macro 'VEC_S16'
     VEC_S16 (21407, 29692, 27969, 25172, 21407, 25172, 27969, 29692);
     ^
idct_altivec.c:62:66: warning: left-hand operand of comma expression has no effect [-Wunused-value]
 #define VEC_S16(a,b,c,d,e,f,g,h) (vector_s16_t) (a, b, c, d, e, f, g, h)
                                                                  ^
idct_altivec.c:72:5: note: in expansion of macro 'VEC_S16'
     VEC_S16 (21407, 29692, 27969, 25172, 21407, 25172, 27969, 29692);
     ^
idct_altivec.c:62:69: warning: left-hand operand of comma expression has no effect [-Wunused-value]
 #define VEC_S16(a,b,c,d,e,f,g,h) (vector_s16_t) (a, b, c, d, e, f, g, h)
                                                                     ^
idct_altivec.c:72:5: note: in expansion of macro 'VEC_S16'
     VEC_S16 (21407, 29692, 27969, 25172, 21407, 25172, 27969, 29692);
     ^
idct_altivec.c:72:5: error: can't convert between vector values of different size
idct_altivec.c:62:51: warning: left-hand operand of comma expression has no effect [-Wunused-value]
 #define VEC_S16(a,b,c,d,e,f,g,h) (vector_s16_t) (a, b, c, d, e, f, g, h)
                                                   ^
idct_altivec.c:74:5: note: in expansion of macro 'VEC_S16'
     VEC_S16 (13623, 18895, 17799, 16019, 13623, 16019, 17799, 18895);
     ^
idct_altivec.c:62:54: warning: left-hand operand of comma expression has no effect [-Wunused-value]
 #define VEC_S16(a,b,c,d,e,f,g,h) (vector_s16_t) (a, b, c, d, e, f, g, h)
                                                      ^
idct_altivec.c:74:5: note: in expansion of macro 'VEC_S16'
     VEC_S16 (13623, 18895, 17799, 16019, 13623, 16019, 17799, 18895);
     ^
idct_altivec.c:62:57: warning: left-hand operand of comma expression has no effect [-Wunused-value]
 #define VEC_S16(a,b,c,d,e,f,g,h) (vector_s16_t) (a, b, c, d, e, f, g, h)
                                                         ^
idct_altivec.c:74:5: note: in expansion of macro 'VEC_S16'
     VEC_S16 (13623, 18895, 17799, 16019, 13623, 16019, 17799, 18895);
     ^
idct_altivec.c:62:60: warning: left-hand operand of comma expression has no effect [-Wunused-value]
 #define VEC_S16(a,b,c,d,e,f,g,h) (vector_s16_t) (a, b, c, d, e, f, g, h)
                                                            ^
idct_altivec.c:74:5: note: in expansion of macro 'VEC_S16'
     VEC_S16 (13623, 18895, 17799, 16019, 13623, 16019, 17799, 18895);
     ^
idct_altivec.c:62:63: warning: left-hand operand of comma expression has no effect [-Wunused-value]
 #define VEC_S16(a,b,c,d,e,f,g,h) (vector_s16_t) (a, b, c, d, e, f, g, h)
                                                               ^
idct_altivec.c:74:5: note: in expansion of macro 'VEC_S16'
     VEC_S16 (13623, 18895, 17799, 16019, 13623, 16019, 17799, 18895);
     ^
idct_altivec.c:62:66: warning: left-hand operand of comma expression has no effect [-Wunused-value]
 #define VEC_S16(a,b,c,d,e,f,g,h) (vector_s16_t) (a, b, c, d, e, f, g, h)
                                                                  ^
idct_altivec.c:74:5: note: in expansion of macro 'VEC_S16'
     VEC_S16 (13623, 18895, 17799, 16019, 13623, 16019, 17799, 18895);
     ^
idct_altivec.c:62:69: warning: left-hand operand of comma expression has no effect [-Wunused-value]
 #define VEC_S16(a,b,c,d,e,f,g,h) (vector_s16_t) (a, b, c, d, e, f, g, h)
                                                                     ^
idct_altivec.c:74:5: note: in expansion of macro 'VEC_S16'
     VEC_S16 (13623, 18895, 17799, 16019, 13623, 16019, 17799, 18895);
     ^
idct_altivec.c:74:5: error: can't convert between vector values of different size
make[5]: *** [libmpeg2arch_la-idct_altivec.lo] Error 1
make[4]: *** [all-recursive] Error 1
make[3]: *** [all-recursive] Error 1
make[2]: *** [all-recursive] Error 1
make[1]: *** [all] Error 2
make: *** [bundle] Error 2

I have applied the same modifications used in the past to allow the OSX ppc compilation but this time with no luck, it seems related to a re-compile of libmpeg2 that should not be needed since I already have the corresponding Macports version installed...

[FrodeSolheim]

You may have to manually tweak FS-UAE's Makefiles to skip building libmpeg2. In recent development versions, there's a configure option for it, but in the older stable/development versions.

[SnakeCoils]

The last 2.7.14dev binary for OSX PPC 10.5.8 is available in The Zone!

About the code tweaking I think I need some hints about that, I have tried to do some rough cut'n'paste from recent development releases but the libmpeg2 still is taken from the FS-UAE archive instead from installed MacPorts one.

[Marmotta]

Originally Posted by SnakeCoils

The last 2.7.14dev binary for OSX PPC 10.5.8 is available in The Zone!

About the code tweaking I think I need some hints about that, I have tried to do some rough cut'n'paste from recent development releases but the libmpeg2 still is taken from the FS-UAE archive instead from installed MacPorts one.

Thanks!!!

The problem with Os4 remain?

[SnakeCoils]

I think the tlosm's patch may still work also on recent development releases, as far as I know it doesn't hurt the normal Amiga 68k behaviour so I could include it in the next dev release.

Hey.
Apparently my account had been deleted... (Posted a patch for uae4arm to build on arm linux a while back....http://eab.abime.net/showpost.php?p=...0&postcount=64 . no clue why my account got deleted??? ) But i did adjust libmpeg using some method, think you need -faltivec and there were some minor change needed, which i cannot find atm, thought i did the work etlegacy. I'll hunt some more, i dont have my ppc powerbook atm..

Edit:
Aah, found it https://github.com/threader/fs-uae/c...010d32a0d09ad7
Edit:
Or -maltivec maybe?
https://github.com/threader/fs-uae/c...1698cc15edc8c2
Implemented a version of the same config i used in e-uae ppc jit while building to identify the cpu. The latest changes i made to update the configure hasnt been merged in the official e-uae repo yet due to issues building on Álmos Rajnai's now dated build system. I spread it around on the mailing lists for tabor and such since i fixed building and needed testing, identifying SPE and correct the asm for the ISA. Speed up somewhat the emulation due to some nit picking the in the code. I can probably dig that up.

Edit.
Added the code from the mail, rename to tar.bz2 .
Issues thats not applied to compile for amigaos , osx and various (?)
"You need to run autoconf and such.
sh bootstrap.sh

add SDL.h to euaeppcjit-code-0/src/sd-sdl/sound.c
--
Just copying some of the issues discussed in the mails with Almos here, it worked on my linux ppc you see.

src/Makefile.am: tools/configure.in tools/configure tools/sysconfig.h.in \
src/tools is changed to .ac

The src/Makefile.am still pointed to the .in file
Line 125 of src/Makefile.am - the offending configure.in:
EXTRA_DIST = \
tools/configure.in tools/configure tools/sysconfig.h.in \
tools/target.h tools/Makefile.in \
test/test_optflag.c test/Makefile.in test/Makefile.am
---
I've seen some oddities when messing with the build system, suddenly it'll fail for no apparent reason.
gcc -o genblitter genblitter.o blitops.o writelog.o
gcc -o build68k build68k.o writelog.o
writelog.o: file not recognized: File truncated
collect2: error: ld returned 1 exit status
Makefile:43: recipe for target 'genblitter' failed
make[2]: *** [genblitter] Error 1

I managed to get this now.
ran a make distclean and tried again

"
Add the patch too against the e-uae-ppc-jit on SF for anyone interested, above still applies. But the changes you want are in configure.ac for fs-uae

Edit: Imported e-uae-ppc-jist to my github and applied the changes and some more. anyone point me to the unofficial ppc osx fs-uae repo too?
"
Can you try building? No ppc atm, and can anyone point me to the guys doing sdl2 for Almos jit?
https://github.com/threader/e-uae-ppc-jit-nix.git
The changes.
https://github.com/threader/e-uae-pp...20b69e4acb8358
"

Originally Posted by SnakeCoils

I think the tlosm's patch may still work also on recent development releases, as far as I know it doesn't hurt the normal Amiga 68k behaviour so I could include it in the next dev release.

https://github.com/threader/fs-uae/b...c/configure.ac
quickly grabbed, merged and added the latest configure changes, just a quickie. git pulled first , but considering forking it off as a fs-uae-ppc? donno if it builds or configs even, can try setting up something in qemu just to check the process.

pear this maybe? but with qemu on ppc only host really works,

/* Set default CPU model based on accelerator board */
if (ISCPUBOARD(BOARD_BLIZZARD, BOARD_BLIZZARD_SUB_PPC)) {
//*model = _T("603ev"); hack hr hr 970gx g5? e500 ? PA6T ?
*model = _T("7450");
} else {
//*model = _T("604e");
*model = _T("host");

Flappy hacking.
https://github.com/threader/fs-uae

[Marmotta]

Originally Posted by SnakeCoils

The last 2.7.14dev binary for OSX PPC 10.5.8 is available in The Zone!

About the code tweaking I think I need some hints about that, I have tried to do some rough cut'n'paste from recent development releases but the libmpeg2 still is taken from the FS-UAE archive instead from installed MacPorts one.

Is compiled with jit and/or altivec?

The performance is very low, with my G5 dual core 2.0 ghz I have only 12 mips (0,65 respect A4000)

[tlosm]

Originally Posted by Marmotta

Is compiled with jit and/or altivec?

The performance is very low, with my G5 dual core 2.0 ghz I have only 12 mips (0,65 respect A4000)

yes lastest are really low in performances compared the before versions.
but are much more accurate.
We have to hope in jit from euae merged by Mike on fsUae

[tlosm]

I think the tlosm's patch may still work also on recent development releases, as far as I know it doesn't hurt the normal Amiga 68k behaviour so I could include it in the next dev release.

yes it work with lastest too. tested on fedora 24 and debian ppc and working.
if some one will fix the endiansess will be better becasue the scsi emulation made thru qemu cyberstorm is more faster in read compared the emulated internal ide

Hola tlosm.

I will not be working on ppc jit for fs-uae this/these rounds, I'm working towards PUAE, and updating/maintaining the old euae, implementing fixes in that version using PUAE as a template. Then hopefully meet in the middle somewhere to then have ppc uae jit ported to PUAE. FS-UAE is written in c++. e and puae is mostly in C. But it took precautions with the code required for all ppc platforms to run ppc jit that say the cache flush routing will compile with c++11. I also made damned sure the implementation is on par with current and other similar functions.

I had to pull the version of e-uae-ppc-jit on my github, malloc problems, i'll work a bit more slowly and do it bit by bit this time, i re-imported the SVN from SF. And i've since learnt and changed say the .gitignore file not to trash the commit so it'll be more readable, a myriad and mermaid got included in the commits as make clean make distclean does not clean up everything, which is another fault , somewhere.

I feel we should use fs-uae for OS4 on Linux PPC and others, and rather use the old e-uae for 68k.

Did you try the fs-uae in my github repo btw? On Linux PPC everything but "host" resulted in a black screen on qemu, used this for testing e-uae-ppc-jit on osx 10.2/3. The goal of that version as i said, is supporting the old platforms which is always done. PUAE also has support for WarpOS.

Using "host" for fs-uae we could leverage the open Sonnet library and run warpup(?) at host speed maybe ?

I'll try and provide a bit more careful update to my e-uae-ppc-jit git repo hopefully today. I'll be away for approx 60 days soon, so i want to push out some code first while i have can.

If your read up on the dev. process of e-uae-ppc-jit changes were needed for G5 to avoid certain instructions which made it horridly slow. These changes are likely be required on fs-uae too. Cant remember the problematic, slow, instruction off the top of my head, but Almos's blog explains it (?)

Originally Posted by tlosm

yes lastest are really low in performances compared the before versions.
but are much more accurate.
We have to hope in jit from euae merged by Mike on fsUae